As per my proposal in

   http://www.cactuscode.org/pipermail/developers/2005-March/000815.html
add a new parameter  copy_ID_after_MoL_PostStep  to control precisely
*when* in CCTK_POSTINITIAL  MoL_FillAllLevels  is scheduled
if  initial_data_is_crap  is set.

The default (MoL_FillAllLevels is scheduled *before* MoL_PostStep)
matches the previous behavior, so there's no change required to par
files unless you want the new behavior (MoL_FillAllLevels is scheduled
*after* MoL_PostStep).


git-svn-id: http://svn.cactuscode.org/arrangements/CactusNumerical/MoL/trunk@87 578cdeb0-5ea1-4b81-8215-5a3b8777ee0b

1 files changed, 15 insertions, 0 deletions

diff --git a/param.ccl b/param.ccl
index d9b6265..d18be54 100644
--- a/param.ccl
+++ b/param.ccl
@@ -116,6 +116,21 @@ BOOLEAN initial_data_is_crap "If the initial data routine fails to set up the pr
 {
 } "no"
 
+# If  initial_data_is_crap  is false, this parameter is ignored.
+# If  initial_data_is_crap  is true, this parameter controls *when*
+#   in CCTK_POSTINITIAL we copy the initial data to previous time levels,
+#   relative to our scheduling of MoL_PostStep.
+#   false ==> copy *before* MoL_PostStep  (default, matches old behavior)
+#   true  ==> copy *after*  MoL_PostStep  (may be preferable for new code)
+BOOLEAN copy_ID_after_MoL_PostStep \
+  "if  initial_data_is_crap  is true, *when* should we copy the		\
+   current time level to all previous time levels:			\
+   false ==> copy *before* MoL_PostStep  (default, matches old behavior)\
+   true  ==> copy *after*  MoL_PostStep  (maybe preferable for new code)"
+{
+} "no"
+
+
 # The default for this parameter corresponds to generic RK2
 STRING Generic_Method_Descriptor "A string used to create a table containing the description of the generic method"
 {